If you’ve ever applied for a job, chances are someone has run a criminal background check on you. But what exactly does that mean?
“There’s often a misconception that you can order a one-stop shop of information on a person, but that doesn’t exist,” says Melissa Sorenson, the executive director of the National Association of Professional Background Screeners.
Instead, a background check typically involves pulling records from multiple places, such as state and county courts. In many cases, these records can be checked online, but some counties haven’t digitized their records or made them available online, meaning a person needs to visit a government office to search the records.
Employers typically farm out this work out to companies that specialize in background checks. It’s a big business. Background screening is a multi-billion-dollar industry, says Rich Wong, a partner at venture-capital firm Accel Partners. He’s betting that Checkr, a startup that closed a $100 million investment this month from Accel and others, will capture a big part of that.
Many of the biggest names in the “gig economy”—including Uber and Instacart—use Checkr’s service to screen workers. But Checkr is also winning fans among more traditional employers. Checkr claims to run one million background checks per month for more than 10,000 companies.
Adecco, one of the world’s largest staffing agencies, uses Checkr in its mobile temporary-work app Adia. Adia CEO Ernesto Lamaina says the company chose Checkr because it’s fast—Checkr says it returns results in 24 hours on average—and could be easily integrated into into the Adia app. “They’re very easy to work with,” Lamaina says.
These two factors, speed and the ability for programmers to easily weave Checkr into an app, are two of its biggest selling points. Instead of a human-resources employee sending a list of names and Social Security numbers to a background-check company, an employer can configure its app to connect with Checkr as soon as a prospective worker applies. The results can be piped right back into the app as well.
Cofounders Jonathan Perichon and Daniel Yanisse came up with the idea for Checkr in 2013 while working as programmers for Deliv, a San Francisco-based on-demand delivery company. They wanted to integrate background screening into the company’s app, but found existing offerings slow, expensive, and designed primarily for large corporations, not small startups. Yanisse says some background screening companies offered services to integrate with their customers’ software, but only to large corporations and government agencies.
The notion of adding tools and services into apps was becoming more commonplace. Twilio, for example, allows developers to add text messaging to an app. Perichon and Yanisse wondered why there wasn’t a similar service for background checks, so they started building their own at night and on weekends; in 2014, they quit their jobs to focus on the idea full-time. “We spent months researching what a background check actually is and how it works,” says Yanisse.
The industry has evolved since 2013. First Advantage, one of the nation’s largest background screening companies, has offered software integration services for organizations of all sizes for three years, CEO Scott Staples says.
Gathering data and documents from multiple agencies and databases is just one step in the process. Once arrest records, court documents, and other files are pulled, someone has to sift through the information to suss out what someone was arrested for, whether they were charged, and whether they were convicted. Perichon points out that because state laws vary, what’s described as felony shoplifting in one place might be treated differently elsewhere.
Checkr attempts to streamline this process by pulling as much information as it can automatically and feeding it into a machine learning system. Documents gathered in-person by contractors are likewise fed into its system. The company works with lawyers to train its algorithms. The result is a system the founders say can summarize disparate documents and identify convictions more quickly than any human.
Once Perichon and Yanisse thought their product was ready, they began emailing gig economy companies in San Francisco and New York. “Uber was at the top of our list,” Perichon says. “I thought ‘maybe we can get them in a couple years.'” But the team didn’t have to wait: someone from Uber reached out to Checkr within the first year. By the end of 2015, Uber was using Checkr to screen all of its US drivers.
For all its success, Checkr isn’t free of controversy. Checkr doesn’t use fingerprints in its background checks, something critics of Uber in particular have called for. San Francisco district attorney George Gascon, for example, famously called background checks that don’t use fingerprints “completely worthless” in 2014.
Perichon and Yanisse acknowledge background checks aren’t perfect. If someone commits a crime away from their home state, a screener could miss a conviction. Records might be missing, outdated, or incorrect. Both humans and machines can misread court documents, or erroneously include convictions for someone with the same name. But the pair argue that machines are better than humans at sifting through documents.
Checkr has faced lawsuits under the Fair Credit Reporting Act, which gives people the right to dispute inaccurate information found in background checks. But spokesman David Patterson says the number of such suits is consistent with other firms conducting a similar number of checks.
Missing someone dangerous is only one problem. Background checks can also penalize good workers for past mistakes.
Activists have persuaded a growing number of companies and governments to not ask about criminal records in job applications. The idea is that hiring managers only run a background check on an employee after deciding that they are qualified, in hopes of reducing bias in the hiring process. But Perichon says companies could do more to eliminate bias, and he thinks Checkr could help, with more-selective reports that include only the most relevant details to an employer. For example, a retailer may not want to know about an old charge for driving under the influence if the job doesn’t involve driving.
Perichon hopes the company can lead by example by hiring people with marred backgrounds. He says nearly 5 percent of the company’s 180 employees have some sort of criminal record, be it a misdemeanor or a felony.